> I'm referencing Jonathan Haidt's work

Why not link to real numbers though? Haidt doesn't understand the numbers, misquotes them out of context, and mangles the data.